human computer interaction cs5340 hci round 2
play

Human-Computer Interaction CS5340 HCI Round 2 Prof. Stephen - PDF document

Human-Computer Interaction CS5340 HCI Round 2 Prof. Stephen Intille http://bit.ly/neu-hci-spring-12 Overview for Today Brief Review of Lecture 1 essentials Brainstorming exercise Review of Homework UI Development Process


  1. Human-Computer Interaction CS5340 – HCI – Round 2 Prof. Stephen Intille http://bit.ly/neu-hci-spring-12 Overview for Today  Brief Review of Lecture 1 essentials  Brainstorming exercise  Review of Homework  UI Development Process  Human Factors  Homework Preview  Research Paper Presentations (Break half way) Overview of Course http://bit.ly/neu-hci-spring-12 1

  2. Administrivia  Stephen  450 WVH, s.intille@neu.edu  Office hours  Most likely Wed early morning (being scheduled)  After class  Send email  Facilitator/Grader - Zeeshan Sayyed  sayyed.z@husky.neu.edu  Class discussion/questions: Piazza http://piazza.com/northeastern/spring2012/cs5340 (Send all questions not specific to your work here) Overview of Course  Texts  Required:  Dix, et al, Human-Computer Interaction  A bit dated, but comprehensive  In bookstore  Other chapters/articles to be provided on Blackboard  Recommended:  Nielsen, Usability Engineering  Norman, The Design of Everyday Things Overview of Course  Weekly Requirements  Read (and absorb!) 50-150 pages  Your reading notes  Individual homework assignment  Team project assignment  Describe and discuss assignments in class  Periodic Requirements  Perform a design session in class  Present a research paper in class 2

  3. Typical Class* Review assignments. Presentation and 1. discussion by randomly selected students Lecture on HCI practice topic 2. Discussion of next week’s assignments 3. Break 4. Intro to research topic by instructor 5. Research paper presentations or design 6. session presentations by students * Changes may be made based on composition of the class Overview of Course  Your reading notes  Bullet lists of most important ideas  Bullet lists of thoughts/ideas generated during reading  Show evidence of thoughtful reading and synthesis of readings throughout course  Post prior to class and hand in hardcopy at class Grading  Prior experience suggests that work in this course will generally fall into one of four categories :  Superior, striking, or unexpected pieces of work with excellent effort demonstrating a mastery of the subject matter and a thoughtful use of concepts discussed in class; work that shows imagination, clarity of presentation, originality, creativity, effort, and attention to detail (A)  Good work demonstrating a capacity to use the subject matter, with adequate preparation and clear presentation (B)  Work that is adequate but that would benefit from increased effort or preparation (C)  Work that needs more effort (D) 3

  4. Breakdown  Your reading notes (10%)  Class presentation(s) (10% )  Individual assignments (30% )  Team assignments (20% ) } 50%  Final project and project presentation (30% ) Schedule  http://bit.ly/neu-hci-spring-12 Overview of Course  Topics covered  HCI theory & practice  A bit on good design  A lot of hands-on experience (You haven’t learned it until you can apply it!)  Cutting-edge HCI research  Topics on your own:  GUI programming in your favorite language  Prerequisites  Programming basics (or see me) 4

  5. Some basic issues & concepts • Ethnography • Task analysis Design • Design guidelines • Scenarios • Expert evaluation • Usability testing Implement Evaluate • Prototyping • GUI tools The HCI development process HCI: listening, adapting, and iteration At every stage! Design Prototype “Typical users” Evaluate Diagram from J. Landay Learn how to observe/listen... 5

  6. Simplify/refine/stress-test tasks  Gotchas  Missing what’s truly important to user  Interruptions  Influence of environment/context  Boredom/lack of novelty  Dealing with problems created by  Environment  Other people  Technological limitations Simplicity is Hard! Some basic issues & concepts Whether the functionality of the system in principle can Utility do what is needed. Usability How well users can use the system’s functionality. + Engagement From Nielsen, Usability Engineering 6

  7. Team Project Major focus of course Will dominate your grade Team Project Guidelines  Your project MUST  Have a substantial UI  Be interactive  Work robustly  Contribute to health or health research  Solve a real-world problem  Be targeted for and tested with older adults Why? Team Project Guidelines  Your project SHOULD  Be creative  Be original  Be non-obvious  Have a “wow” factor  Allow you, at the end of this course, to leapfrog your peers with an amazing demo! Why? 7

  8. Team Project Constraints  Team: 3-4 members, ideally multi- disciplinary  Focus: Health Application for (or used by) older adult users  Context: Senior center, home, etc.  Platform: Your choosing  Input/output/sensing: Your choosing Team Project Categories  App for older adults in senior center (to facilitate goals/tasks you identify)  “Serious game” for older adults to generate food nutrition database  App for older adults that meets guidelines for an available app competition (e.g., http://www.health2challenge.org/healthy-people-2020-leading-health-indicators-app-challenge/ ) (caveats) Team Project Brainstorming Exercise 8

  9. Project Brainstorming  Think about a graphical user interface you’d like to build  Should be representative of your interests  No commitment  Sketch out the idea  Put a title and your name on it  Be ready to talk about it  15 minutes Individual Homework # 1 UI Critique  Find 2 good & 2 bad examples of UI design  Some criteria  Consistency (inter & intra application)  Prevent errors  Permit error correction  Obviousness (“affordances”)  Feedback  KISS  Include visuals if possible Overview Dix Forward and Introduction 9

  10. Important take-aways  HCI is difficult, rewarding, necessary  Multi-disciplinary  Trying to get at scientifically rigorous ways to “know thy user”  Errors result from “narrow optimization” that fails to account for context (especially human kind) Important take-aways  Usability analysis is nice .. But too late. Design is where the action is.  Those who can evaluate but not design at a disadvantage!  HCI is a discipline that outlines processes to help you with a very difficult task X-centered design  What do we want X to be and why? 10

  11. Don’t do this...  “They’ll do [x]” example  Fingernail example  “Don’t have time” example Do this...  Spend the time HCI requires  Learn to listen  And listen some more  And listen some more  Trust the process  Have courage: throw out ideas that are not working Components  People  Computers  Tasks  Usability  Useful  Usable  Used  “Inventive inspiration”  Tricks of trade (e.g. architects) 11

  12. Human Factors (the people) Dix Ch 1 Human Factors  A body of scientific facts about human capabilities and limitations.  The study of how humans behave physically and psychologically in relation to particular environments, products, or services.  aka Ergonomics Human Factors Highlights  Inputs  Visual  Auditory  Haptic  Olfactory  Vestibular  Outputs  Motor (hands, feet, head, gaze, speech, …)  Neural 12

  13. Haptic Olfactory Jofish Kaye Vestibular 13

  14. Head Gaze Gaze tracking Vision  Why can’t you use color alone as an output modality?  8% males and 1% females color blind 14

  15. Vision  How can you tell if your display will suffer from optical illusions, or cause users to become dizzy or nauseous?  Test it with real users! Visual issues impact design  Examples  Magnify horizontal lines and reduce vertical (to look square, must be slightly tall)  Optical center – we see center of page as a little above optical center  WHY IS THIS HARDER TO READ?  Contrast  Negative (preferred but watch for flicker)  Positive 15

  16. Blanking 16

  17. Blanking Eye movement [Yarbus] 17

  18. You don’t “see” what you think Scene Mental model You don’t “see” what you think What if something changes here? Blanking 18

  19. Blanking 19

  20. OSBA need change reasoning system User model Activity Context Application 1 Application 2 recognition Object Application n Change positions Sort display requests reasoning Sensor data •Immediate msg User 1 •Immediate msg only User 2 Display Sensor •Info update wall display (facing) •Info update when convenient controller input … Object 4 Table display Object 3 Object 1 Wall display User 1 Object 2 User 2 PDA display Reasoning  We fill in the gaps  This can lead to faulty mental models  Novices  Group problems according to superficial characteristics  Experts  Group problems based on underlying conceptual similarities Errors  Two major classes  Changes in the context of skilled behavior  Familiar overrides unfamiliar  Incorrect understanding of a situation (“mental models”)  E.g., CITY postpone  Variation from conventions 20

Recommend


More recommend